<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/">
  <channel>
    <title>Swift on wid@internet:~$ </title>
    <link>https://widnyana.web.id/en/tags/swift/</link>
    <description>Recent content in Swift on wid@internet:~$ </description>
    <generator>Hugo</generator>
    <language>en-us</language>
    <lastBuildDate>Wed, 11 Nov 2026 19:04:02 +0700</lastBuildDate>
    <atom:link href="https://widnyana.web.id/en/tags/swift/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Perceptual hash from scratch: 32x32 pixels, one DCT, 64 bits, in Swift</title>
      <link>https://widnyana.web.id/en/posts/perceptual-hash-dct-explained/</link>
      <pubDate>Mon, 09 Nov 2026 19:00:00 +0700</pubDate>
      <guid>https://widnyana.web.id/en/posts/perceptual-hash-dct-explained/</guid>
      <description>A normal hash changes completely when one pixel changes, so it cannot find near-duplicate photos. A perceptual hash shrinks the image to 32x32 grayscale, runs a discrete cosine transform, keeps the 64 lowest frequencies, and writes one bit per coefficient. Full Swift reference, real output, and where it breaks.</description>
    </item>
  </channel>
</rss>
